\begin{table}[h!]
    \centering
    \caption{WTP and Cost by Quartile of MVPF Distribution}
    \label{tab:quartiles_mvpf}
    \begin{tabular*}{\textwidth}{@{\extracolsep{\fill}}lcccc@{}}
\toprule 
           & \multicolumn{4}{c}{Quartile of Neighborhood MVPF Distribution}     \\ \cmidrule{2-5}
           & 1            & 2            & 3           & 4           \\ \midrule \addlinespace
WTP        & \$$$wtp_q1_est$$  & \$$$wtp_q2_est$$  & \$$$wtp_q3_est$$ & \$$$wtp_q4_est$$ \\
           & (\$$$wtp_q1_se$$)  & (\$$$wtp_q2_se$$)  & (\$$$wtp_q3_se$$) & (\$$$wtp_q4_se$$) \\ \addlinespace
Net Cost   & \$$$cost_q1_est$$  & \$$$cost_q2_est$$  & \$$$cost_q3_est$$ & \$$$cost_q4_est$$ \\
           & (\$$$cost_q1_se$$)  & (\$$$cost_q2_se$$)  & (\$$$cost_q3_se$$) & (\$$$cost_q4_se$$) \\  \addlinespace
Gross Cost & \$$$gcost_q1_est$$  & \$$$gcost_q2_est$$  & \$$$gcost_q3_est$$ & \$$$gcost_q4_est$$ \\
           & (\$$$gcost_q1_se$$)  & (\$$$gcost_q2_se$$)  & (\$$$gcost_q3_se$$) & (\$$$gcost_q4_se$$) \\ \addlinespace \bottomrule
\end{tabular*}
	\begin{spacing}{1}
\begin{tablenotes}
      \item \footnotesize \textit{Notes:} This table reports, by quartile of the distribution of neighborhoods ranked by the estimated marginal value of public funds (MVPF) for 421-a, the willingness to pay (WTP) and net and gross cost (net and cost of fiscal externalities). All values are computed on a per-household, not per-unit, basis (see Appendix \ref{appendix:extensions}). Neighborhoods are ranked and grouped once on their actual MVPF estimate, and standard errors are computed by a cluster-bootstrap at the level of Neighborhood Tabulation Areas. Therefore, the standard errors incorporate three sources of sampling variance: reported standard errors from the Opportunity Atlas mobility estimates \citep{chetty2018opportunity}, standard errors in my estimates of marginal cost per unit, and misclassification of neighborhoods into quartiles due to uncertainty in the estimated MVPFs. This analysis is therefore responsive to concerns in \citetappendix{mogstad2020inference} about the implications of sampling variance for the analysis of Moving to Opportunity and similar policies. $\sym{*} = p < 0.10$, $\sym{**} = p < 0.05$, $\sym{***} = p < 0.01$.
      \end{tablenotes}
      \end{spacing}
\end{table}
